Systematic uncertainties in heavy ion collision centrality measures

The collision centrality for relativistic heavy-ion scattering events is often reported in terms of geometrical quantities using a model to relate the latter to multiplicity [1]. Using a Monte Carlo Glauber model the accuracy of several geometrical measures is reported for collision systems relevant to the RHIC program. The measures include impact parameter, number of interacting nucleons ($N_{part}$), number of binary interactions ($N_{bin}$), and the average number of binary collisions per incident participant nucleon $\nu = N_{bin}/(N_{part}/2)$ for Au-Au collisions at $\sqrt{s_{NN}}$ = 20, 62, 130 and 200 GeV and Cu-Cu at 62 and 200 GeV. Systematic uncertainties in the centrality measures due to errors in the matter densities, nucleon-nucleon cross section, multiplicity production model, and measured multiplicity frequency distribution are estimated. We find that the impact parameter is most accurately determined, followed closely by $\nu$. Centrality measures $N_{part}$ and $N_{bin}$ can be significantly affected by experimental uncertainties in the multiplicity frequency distribution [1], particularly that caused by trigger and primary vertex finding inefficiencies for low multiplicity events. Combined systematic errors for each collision system are given. \newline [1] C. Adler, et al., Phys. Rev. Lett. 87, 112303 (2001).